RITHM Dataset : Robotic Inductive Tasks to monitor Human Mental state

This dataset contains multimodal recordings collected during a human–robot interaction experiment. Thirty participants took part in two experimental sessions, each involving two interaction tasks. The dataset includes continuous physiological recordings (EEG, ECG, EDA, PPG), behavioral performance measures (response times, accuracy), and subjective ratings of mental effort. Data were originally acquired using the Lab Streaming Layer (LSL) framework and recorded with LabRecorder. The recordings were subsequently converted to Brain Imaging Data Structure (BIDS)–compatible formats. EEG data are provided in EDF format, while peripheral physiological signals are provided in both EDF and tab-separated value (TSV) formats. Event markers corresponding to stimuli and participant responses are provided as TSV files. The dataset is organized according to the BIDS specification and contains continuous, non-epoched recordings without filtering or artifact rejection. While BIDS standards are primarily defined for neural data (e.g., EEG), the same organizational principles were consistently extended to peripheral physiological signals to ensure coherence across modalities. This dataset is intended to support research on cognitive effort during human–robot interaction, using both modality-specific and multimodal signal analyses. It is particularly suited for investigating inter-individual variability as well as differences across sessions and tasks.